A WeeChat native client for Slack.com. Provides supplemental features only available in the web/mobile clients such as: synchronizing read markers, typing notification, threads (and more)! Connects via the Slack API, and maintains a persistent websocket for notification of events.
- Features
- Contributing
- Dependencies
- Setup
- Commands and options
- Removing a team
- Optional settings
- FAQ
- Known issues
- Debugging
- Support
- Threads support
- Slack status support
- Slash commands (including custom ones)
- Upload to slack capabilities
- Emoji reactions
- Edited messages work just like the official clients, where the original message changes and has (edited) appended.
- Unfurled urls dont generate a new message, but replace the original with more info as it is received.
- Regex message editing (s/oldtext/newtext/)
- Smarter redraw of dynamic buffer info (much lower CPU %)
- Multiple Teams supported. Just add multiple api tokens separated by commas
- Replays history automatically during startup. (and sets read marker to the correct position in history)
- Open channels synchronized with Slack. When you open/close a channel on another client it is reflected in wee-slack
- Colorized nicks in chat
- Supports bidirectional slack read notifications for all channels. (never reread the same messages on the web client or other devices).
- Typing notification, so you can see when others are typing, and they can see when you type. Appears globally for direct messages
- Away/back status handling
- Expands/shows metadata for things like tweets/links
- Super fun debug mode. See what the websocket is saying
See docs/contributing.md.
- WeeChat 1.3+ http://weechat.org/
- websocket-client https://pypi.python.org/pypi/websocket-client/
- Since WeeChat 2.6, Python 3 modules are required, see https://weechat.org/blog/post/2019/07/02/Python-3-by-default
- Some distributions package WeeChat's plugin functionalities in separate packages.
Be sure that your WeeChat supports python plugins. Under Debian, install
weechat-python
Arch Linux: pacman -S python-websocket-client
Debian/Ubuntu: apt install weechat-python python3-websocket
Fedora: dnf install python3-websocket-client
FreeBSD: pkg install py36-websocket-client
OpenBSD: pkg_add weechat-python py3-websocket-client
Other: python3 -m pip install websocket-client
The easiest way to install wee-slack is from inside WeeChat by running:
/script install slack.py
This will install and load it and enable automatically loading it when WeeChat starts.
Alternatively, you can download the wee_slack.py
script from the GitHub repo,
either from the latest release or from the master branch and place it under the
python/
directory in the WeeChat data files
directory.
If you use the default XDG directories, this will be
$XDG_DATA_HOME/weechat/python
, or ~/.local/share/weechat/python
if
$XDG_DATA_HOME
is not set. If you use the old single directory, it will be
~/.weechat/python
.
If you use this method and want it to load automatically when WeeChat starts,
you have to create a symlink from the autoload
directory to the script.
E.g. to install and automatically load the latest version from the master
branch when using ~/.local/share/weechat
:
mkdir -p ~/.local/share/weechat/python/autoload
cd ~/.local/share/weechat/python
curl -O https://raw.githubusercontent.com/wee-slack/wee-slack/master/wee_slack.py
ln -s ../wee_slack.py autoload
If you already had WeeChat running when installing the script, or if you didn't
enable autoload, you have to load the
script
by running this inside WeeChat: /script load wee_slack.py
There are two types of tokens that can be used, OAuth tokens and session tokens. The official way to get a token is to use OAuth. However, this has several drawbacks, so an alternative way is to pull a session token out of the web client.
Drawbacks of OAuth tokens:
- If the team is restricting app installations, wee-slack has to be approved by an admin.
- For free teams, wee-slack will use one of the ten app slots.
- The subscribe and unsubscribe commands won't work.
- Threads can only be marked as read locally, it won't sync to Slack. This means they will be unread again after reloading the script.
Drawbacks of session tokens:
- These tokens can't be revoked, so be careful not to loose them.
- They are not officially supported, and may stop working at any time.
Log in to Slack:
/slack register
This command prints a link you should open in your browser to authorize WeeChat with Slack. If the page shows a different team than the one you want to add, you can change the team in the top right corner of the page.
Once you've accomplished this, the page will show a command which you should run in WeeChat. The command is of the form:
/slack register <code>
Note that this will store your token in plain text in the
plugins.var.python.slack.slack_api_token
option. See "Optional: Secure the
tokens" below for how to secure it.
Your Slack team is now added, and you can complete the setup by reloading the wee-slack script.
/python reload slack
Note that by default GitHub Pages will see a temporary code used to create your
token (but not the token itself). If you're worried about this, you can use the
-nothirdparty
option, though the process will be a bit less user friendly.
- Open and sign into the Slack customization page. Check that you end up on the correct team.
- Open the developer console (
Ctrl+Shift+J
/Cmd+Opt+J
in Chrome andCtrl+Shift+K
/Cmd+Opt+K
in Firefox). - Paste and run this code:
window.prompt("Session token:", TS.boot_data.api_token)
- A prompt with the token will appear. Copy the token.
- In the developer console go to Application in Chrome or Storage in Firefox.
- Expand Cookies and click on the domain.
- Find the cookie named
d
and copy the value. - Return to WeeChat and run
/slack register <token>:<cookie>
. - Reload the script with
/python reload slack
.
Note that this will store your token in plain text in the
plugins.var.python.slack.slack_api_token
option. See "Optional: Secure the
tokens" below for how to secure it.
Note that if you log in or out of any teams in your browser, the cookie will be invalidated, and you will have to update it.
Note that in some cases it may be necessary to include the d-s
cookie as
well. If so, you can supply it in this format <token>:d=<d_cookie>;d-s=<d-s_cookie>
.
If you use Chrome or Firefox, you can run the extract_token_from_browser.py
script to get the tokens and cookies for all the teams you're logged into:
./extract_token_from_browser.py <browser>
(Note this script requires the Plyvel, PyCryptodome and SecretStorage libraries for Chrome and may require the python-snappy library for Firefox.)
You can run the register command multiple times to connect to multiple teams. If you set the token option yourself, you should separate the tokens with commas.
/mute /set plugins.var.python.slack.slack_api_token <token1>,<token2>,<token3>
The /mute
prefix is to prevent "Option changed ..." from being printed to the
core buffer, so your token is not logged in your buffer logs.
The tokens you add will be stored as plain text in the option
plugins.var.python.slack.slack_api_token
. If you don't want to store your API
token in plain text you can use the secure features of WeeChat:
/secure passphrase this is a super secret password
/secure set slack_token <YOUR_SLACK_TOKEN>
/set plugins.var.python.slack.slack_api_token ${sec.data.slack_token}
Note that you will have to move your tokens manually from
plugins.var.python.slack.slack_api_token
to the secure variable after each
time you run /slack register <code>
.
For the available options see docs/Options.md or run this command:
/set slack
Most options require that you reload the script with /python reload slack
after changing it to take effect.
For the available commands see docs/Commands.md or run this command:
/slack help
In addition to the commands listed with /slack help
, most normal IRC
commands, like /join
, /part
, /query
, /msg
, /me
, /topic
, /away
and
/whois
work normally. See WeeChat's
documentation or
/help <cmd>
if you are unfamiliar with these.
There are also some special messages you can send:
Modify previous message using regex:
s/old text/new text/
Modify 3rd previous message using regex:
3s/old text/new text/
The regex also supports the flags g
for replacing all instances, i
for
ignoring case, m
for making ^
and $
match the start/end of each line and
s
for making .
match a newline too. Use them by appending one or more of
them to the regex:
s/old text/new text/gi
Delete previous message:
s///
Add a reaction to the nth last message. The number can be omitted and defaults to the last message. The +
can be replaced with a -
to remove a reaction instead.
3+:smile:
To send a command as a normal message instead of performing the action, prefix it with a slash or a space, like so:
//slack
s/a/b/
Start a new thread on the most recent message The number indicates which message in the buffer to reply to, in reverse time order:
/reply 1 here is a threaded reply to the most recent message!
Open an existing thread as a channel. The argument is the thread identifier, which is printed in square brackets with every threaded message in a channel:
/thread af8
To access the last thread in a channel a shorthand is available:
/thread
Label a thread with a memorable name. The above command will open a channel called af8, but perhaps you want to call it "meetingnotes". To do so, select that buffer and type:
/label meetingnotes
Note: labels do not persist once a thread buffer is closed
To enable rendering of emoji characters and tab completion of emoji names, copy
or symlink the
weemoji.json
file to your WeeChat config directory (e.g. ~/.local/share/weechat
or
~/.weechat
). If doing this after starting wee-slack, you will have to reload
it by running /python reload slack
.
Emoji names can be completed by typing colon and the start of the emoji name and pressing tab.
The cursor mode and mouse mode can be used to interact with older messages, for editing, deleting, reacting and replying to a message. Mouse mode can be toggled by pressing Alt
+m
and cursor mode can be entered by running /cursor
(see /help cursor
).
If mouse mode is enabled, the default behavior when right-clicking on a message is to paste its id in the input. It can be used in /reply
, s/
substitution/deletion and in +:emoji:
commands instead of a message number.
It can also be used as an argument to the /slack linkarchive
command.
In cursor mode, the M
key achieves the same result (memo: the default for WeeChat is to paste the message with m
, M
simply copies the id).
In addition, R
will prepare a /reply id
and D
will delete the message (provided it's yours).
T
will open the thread associated to a message, equivalent to /thread id
L
will call the /slack linkarchive
command behind the hood and paste it to the current input.
Please see WeeChat's documentation about how to use the cursor mode or adapt the bindings to your preference.
Default key bindings:
/key bindctxt mouse @chat(python.*):button2 hsignal:slack_mouse
/key bindctxt cursor @chat(python.*):D hsignal:slack_cursor_delete
/key bindctxt cursor @chat(python.*):L hsignal:slack_cursor_linkarchive
/key bindctxt cursor @chat(python.*):M hsignal:slack_cursor_message
/key bindctxt cursor @chat(python.*):R hsignal:slack_cursor_reply
/key bindctxt cursor @chat(python.*):T hsignal:slack_cursor_thread
Note that if these keys are already defined, they will not be overwritten by wee-slack. In that case, you will have to define your own key bindings by running the above commands modified to your liking.
hsignals slack_mouse
and slack_cursor_message
currently have the same meaning but may be subject to evolutions.
You may remove a team by removing its token from the dedicated comma-separated list:
/set plugins.var.python.slack.slack_api_token "xoxp-XXXXXXXX,xoxp-XXXXXXXX"
You can use tab completion after the key to complete the current value. To see
which token belongs to which team, run /slack teams
.
After removing the token, you have to reload wee-slack with /python reload slack
.
Show typing notification in main bar (slack_typing_notice):
/set weechat.bar.status.items [buffer_count],[buffer_plugin],buffer_number+:+buffer_name+{buffer_nicklist_count}+buffer_filter,[hotlist],completion,scroll,slack_typing_notice
Show channel name in hotlist after activity
/set weechat.look.hotlist_names_level 14
Install the script
autosort.py by running
/script install autosort.py
. This will keep your buffer list sorted
alphabetically by default. If you want to customize it, run /help autosort
.
Run /set irc.look.server_buffer independent
and install the
autosort.py script
mentioned in the previous question.
Use this trigger.
You need the notify-send
command, or alternatively replace it with another
command in the trigger.
Use the notification_center.py script. You can install it with /script install notification_center.py
.
There are many scripts in the scripts repo for various use cases. Note that not all may work with wee-slack, so you will have to test them.
With WeeChat >= 4.0.0 this is supported out of the box. You can either press Alt+Enter to insert a line break, or paste something containing multiple lines.
For older versions of WeeChat, continue reading the next section.
You have to install a script to be able to send multiple lines, e.g. the
multiline.pl
script with: /script install multiline.pl
By default it will wait for one second after you press enter, and if you type another character in that period, it will insert the character on a newline, and if you don't type anything it will send the message. If you rather want to use a separate key to insert a newline, and have the enter key send the message immediately, you can run these commands:
/set plugins.var.perl.multiline.magic_paste_only on
/key bind meta-ctrl-M /input insert \n
This will bind meta-enter (which is usually alt-enter) to insert the newline.
Replace meta-ctrl-M
with something else if you want to use a different key
combination.
The multiline.pl
script will also let you edit pasted text which incudes
newlines before you send the message. If this is not working, you may try to
run the commands below. At least in the kitty
terminal, it won't work by
default, but should work after running these commands:
/set plugins.var.perl.multiline.weechat_paste_fix "off"
/key bind ctrl-J /input magic_enter
You may also want to disable WeeChat's paste prompt, since that is not necessary
when using multiline.pl
:
/set weechat.look.paste_max_lines -1
Not all issues are listed here (see issues for all), but these are some noteworthy:
- If you set
background_load_all_history
tofalse
:- Channels will not be shown as unread when wee-slack loads, even if there are unread messages. Messages which arrive after wee-slack has loaded however will mark the channel as unread.
- If messages arrive while the connection to Slack is lost (e.g. during suspend), they will not appear in the hotlist.
- If you use an OAuth token or a legacy token instead of a session token:
- Threads can only be marked as read locally, it won't sync to Slack. This means they will be unread again after reloading the script.
To help debugging you can enable debugging output about what wee-slack is doing
by enabling debug mode and changing debug level (between 0 and 5, default is 3,
decrease to increase logging and vice versa). Enabling this will open a new
buffer slack-debug
where the messages are printed. Enable it and change level
by running:
/set plugins.var.python.slack.debug_mode on
/set plugins.var.python.slack.debug_level 0
You can also dump all the JSON responses received from the API in
/tmp/weeslack-debug/
. This requires a script reload after enabling. Enable it
with:
/set plugins.var.python.slack.record_events true
/python reload slack
wee-slack is provided without any warranty whatsoever, but you are welcome to ask questions in #wee-slack on Libera.Chat.